THE INTERPLAY OF EXTINCTION AND SYNCHRONY IN THE DYNAMICS OF METAPOPULATION FORMATION John Vandermeer and Zachary Hajian-Forooshani Department of Ecology and Evolutionary Biology University of Michigan Ann Arbor, MI 48109 Corresponding author. J. Vandermeer [email protected] Abstract The idea of a metapopulation has become canonical in ecology. Its original mean field form provides the important intuition that migration and extinction interact to determine the dynamics of a population composed of subpopulations. From its conception, it has been evident that the very essence of the metapopulation paradigm centers on the process of local extinction. We note that there are two qualitatively distinct types of extinction, gradual and catastrophic, and explore their impact on the dynamics of metapopulation formation using discrete iterative maps. First, by modifying the classic logistic map with the addition of the Allee effect, we show that catastrophic local extinctions in subpopulations are a pre-requisite of metapopulation formation. When subpopulations experience gradual extinction, increased migration rates force synchrony and drive the metapopulation below the Allee point resulting in migration induced destabilization of the system across parameter space. Second, a sawtooth map (an extension of the Bernoulli bit shift map) is employed to simultaneously explore the increasing and decreasing modes of population behavior. We conclude with four generalizations. 1. At low migration rates, a metapopulation may go extinct faster than completely unconnected subpopulations. 2. There exists a gradient between stable metapopulation formation and population synchrony, with critical transitions from no metapopulation to metapopulation to synchronization, the latter frequently inducing metapopulation extinction. 3. Synchronization patterns emerge through time, resulting in synchrony groups and chimeric populations existing simultaneously. 4. There are two distinct mechanisms of synchronization: i. extinction and rescue and, ii.) stretch reversals in a modification of the classic chaotic stretching and folding. 1 Contents 1. Introduction 2 2. The logic and dynamics of the logistic/Allee map 5 2.1 - The emergence of a metapopulation with catastrophic extinction 7 2.2 - The subcritical state of the logistic/Allee map 15 3. The ecological significance of the sawtooth map 16 3.1 - Local extinction and synchrony patterns from the sawtooth map 20 3.2 Metapopulation collapse at low migration 23 3.3 The dual nature of synchrony and the “metapopulation window” 25 4. Discussion 29 1. Introduction. There is a fundamental contradiction in metapopulation theory. On the one hand, general qualitative understanding assumes that local isolated populations tend to go extinct but that if they are interconnected by migration, a collection of such “subpopulations” could persist indefinitely, the very meaning of metapopulation. On the other hand, a substantial literature notes that if populations are coupled, repeated low points in population densities may become synchroniZed so that if, at a particular point in time, an extinction force visits one such population, it will visit all -- i.e., extinction of the entire collection of subpopulations (the metapopulation) is expected (Fox et al., 2017; Wang et al., 2015). It may seem that metapopulation theory thus must conclude that interpopulation migration is “stabiliZing,” but only to some level of migration, after which it is “destabiliZing,” the two forces apparently in opposition, but interacting in sometimes complicated ways (Abbott, 2011). A successful metapopulation thus must strike this natural balance (Griffen, and Drake, 2008). Herein we demonstrate that, at least through the lens of discrete iterative single species equations, complications may arise associated first, with the form of extinction and second, with the dynamics of synchrony. Extinction of local subpopulations is a core idea of metapopulation theory, whether in its original mean field form (Levins, 1969) or in the many enriching embellishments subsequently proffered (e.g., Hanski 1998; Gilpin 2012). It remains core to the very idea of a metapopulation, providing the important intuition that migration and extinction combine to determine whether a population persists or not. The evident connection with the Allee effect has also been noted (Amarasekare, 1998). Although empiricists sometimes view the Allee effect as rare in nature (e.g. Gregory et al., 2010), from a theoretical point of view local extinction of subpopulations does not occur without an Allee point (allowing an Allee point of Zero is an obvious option to maintain generality). And, of course, if local extinctions of subpopulations do not occur, metapopulations do not exist. There is an extensive literature dealing with the extinction process. Two approaches seem evident; first, the obvious idea that stochastic forces are likely to result in extinction especially in rare populations (Lande, 1993), and second, extinction emerges from dynamic and deterministic population forces (Schreiber, 2003). Deterministic population extinction emerges in either of the classical predator prey models (i.e. Lotka/Volterra or Nicholson/Bailey), and it seems to be tacitly assumed that single population models produce extinction only when population growth rate is less then 1.0 (Gaggiotti and Hanski 2004). Yet with a combination of the Allee effect and either chaotic or intermittent populations in a discrete time framework extinction occurs with an obvious 2 mechanism -- if the minimum population siZe in a chaotic (or intermittent) attractor is less than the Allee point, the population will eventually go extinct (Schreiber, 2003; Vandermeer, 2020; 2021) with no stochastic force needed. A rich theoretical literature is relevant, acknowledging that the Allee point is actually a separatrix of the basin of attraction of the Zero point, and that extinction is thus inevitable if the attractor includes reachable points below that separatrix (Grebogi et al., 1982; 1983; Vandermeer and Yodzis, 1999; Zotos et al., 2021). Here, we explore this general assumption. We note that if extinction is to arise from deterministic forces, we must assume either 1) all subpopulations are declining all the time or 2) some form of intermittent behavior is exhibited by the subpopulations such that an extinction threshold (either Zero or an Allee critical point) is expected to be breached at some time. The idea of population synchrony has likewise become conventional wisdom in ecology. From elementary considerations of classical equations (Vandermeer, 1993; 2006) to more thoughtful considerations of ecological interactions in general (Platt and Denman, 1975), ecological populations under a variety of circumstances behave like other oscillators in nature -- they form synchrony patterns (StrogatZ, 2012). In practice, numerous cases of phase locking have been reported from natural populations (Benincá et al., 2009; Blasius, et. al., 1999; Blasius and Stone, 2000a; Earn et al., 1998) and a variety of theoretical formulations reinforce the basic idea ( Koelle and Vandermeer, 2005; Goldwyn and Hastings, 2008; Nobel et al., 2015; Ahn and Rubchinsky, 2020; Azizi and Kerr, 2020). We consider a metapopulation as a collection of “propagating sinks” to use a category from the body of literature generally called source/sink populations (Pulliam, 1988; Vandermeer et al., 2010). By definition each subpopulation is doomed to local extinction, but sends out propagules before the extinction sets in. A metapopulation is thus a collection of propagating sinks. Intuitively, the Levins result notes that extinction rates must be smaller than migration rates for the whole metapopulation to persist. Each subpopulation (propagating sink) is destined to local extinction, but if migration among subpopulations is larger than extinction, the collection of subpopulations may persist in perpetuity. Yet, the coupling imposed by the inter-subpopulation migration also implies synchrony of populations, which implies eventual extinction of the whole metapopulation. The extinction process is well-appreciated as containing a variety of complicating issues. Nevertheless, much of the literature seems to follow Darwin’s simple observation that “Rarity, as geology tells us, is the precursor of extinction” (Williamson, 1989). While Darwin and his successors emphasiZe the extinction of entire clades, ecological dynamics are concerned with local extinctions, relevant to larger issues such as island biogeography (MacArthur and Wilson, 1967; Losos and Ricklefs, 2009), conservation (Lande, 1998), and, most importantly for this paper, metapopulations. In all applications, extinction is frequently tied in, perhaps only tacitly, with the idea of being rare (Hartley and Kunin, 2003). One gains concern in conservation, for example, when a species is thought to be rare enough to pass some lower threshold and thus be in danger of extinction. RabinowitZ’s (1981) classic framework of seven types of rarity, grew from the fundamental insight that rareness happens for a variety of reasons and categoriZing those reasons could aid further research into the topic. Much subsequent attention to rarity had (and still has) to do with the implicit assumption that populations headed for extinction are likely to be rare (Harnik, et al., 2012). As repeatedly noted, both empirically and theoretically, this assumption is not necessarily true (e.g., Wayne et al., 2015). Yet it is generally thought